ChatGPT-4 Vision Struggles with Radiologic Image Interpretation
ChatGPT-4 Vision shows strong performance on text-based radiology questions but struggles with image interpretation, highlighting limitations in its current application to radiology.
Category: Cloud Computing
Cloud Computing
Latest
Cloud Computing
Latest
AI in Imaging: Kids and Teens See Promise, Stress Need for Supervision
Children and young people are generally positive about AI in healthcare, especially in radiology, but emphasize the need for human supervision, according to a U.K.-based survey.
Cloud Computing
Latest
Ransomware Threats in GE’s Vivid Ultrasound
Nozomi Networks found 11 vulnerabilities in GE HealthCare's Vivid Ultrasound systems, risking ransomware and data manipulation. Physical access is needed to exploit these flaws, and patches are available on GE HealthCare's Product Security Portal.
New AI Method Captures Uncertainty in Medical Images
A new AI tool named Tyche addresses the ambiguity in medical image segmentation by providing multiple plausible segmentations without the need for retraining, potentially advancing diagnostic accuracy and biomedical research.
Read More
AI Boosts Breast Cancer Screening Accuracy
Combining AI with radiologists’ assessments may enhance breast cancer screening with an algorithm tailored to identify normal mammograms accurately, potentially reducing false positives and unnecessary follow-ups while maintaining cancer detection rates.
Read More
Researchers Develop Deep Learning Model to Predict Breast Cancer
Researchers have developed AsymMirai, an interpretable AI model using mammograms to predict breast cancer risk one to five years in advance by comparing differences in left and right breast tissue, showing similar performance to the complex Mirai model and suggesting bilateral dissimilarity as a potential marker for breast cancer risk.
Read More
How AI Tools Impact Radiologists Differently
Recent research suggests that while medical AI tools have the potential to enhance the interpretation of images like X-rays and CT scans for more accurate diagnoses, their effectiveness may differ among clinicians.
Read More
Researchers Look at Environmental Impacts of AI Tools
As artificial intelligence (AI) is increasingly used in radiology, researchers caution that it’s essential to consider the environmental impact of AI tools.
Read More